WIP: onnx support

lambda_diffusers/__init__.py

@@ -1 +1,2 @@
-from .pipelines import StableDiffusionImageEmbedPipeline
+from .pipelines import StableDiffusionImageEmbedPipeline
+from .pipelines import StableDiffusionImageEmbedOnnxPipeline

lambda_diffusers/pipelines/__init__.py

@@ -1 +1,3 @@
-from .pipeline_stable_diffusion_im_embed import StableDiffusionImageEmbedPipeline
+from .pipeline_stable_diffusion_im_embed import StableDiffusionImageEmbedPipeline
+from .pipeline_stable_diffusion_im_embed_onnx import StableDiffusionImageEmbedOnnxPipeline
+

lambda_diffusers/pipelines/pipeline_stable_diffusion_im_embed_onnx.py

@@ -0,0 +1,169 @@
+import inspect
+from typing import List, Optional, Union
+
+import numpy as np
+
+from transformers import CLIPModel, CLIPFeatureExtractor, CLIPTokenizer, CLIPVisionModel
+
+from diffusers.onnx_utils import OnnxRuntimeModel
+from diffusers.pipeline_utils import DiffusionPipeline
+from diffusers.schedulers import DDIMScheduler, LMSDiscreteScheduler, PNDMScheduler
+from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput
+from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
+
+import torch
+import PIL
+import warnings
+
+
+class StableDiffusionImageEmbedOnnxPipeline(DiffusionPipeline):
+    vae_decoder: OnnxRuntimeModel
+    unet: OnnxRuntimeModel
+    scheduler: Union[DDIMScheduler, PNDMScheduler, LMSDiscreteScheduler]
+    safety_checker: StableDiffusionSafetyChecker
+
+    def __init__(
+        self,
+        vae_decoder: OnnxRuntimeModel,
+        unet: OnnxRuntimeModel,
+        scheduler: Union[DDIMScheduler, PNDMScheduler, LMSDiscreteScheduler],
+    ):
+        super().__init__()
+        scheduler = scheduler.set_format("np")
+        self.register_modules(
+            vae_decoder=vae_decoder,
+            unet=unet,
+            scheduler=scheduler,
+        )
+        self.feature_extractor = CLIPFeatureExtractor.from_pretrained("openai/clip-vit-large-patch14")
+        self.image_encoder = CLIPModel.from_pretrained("openai/clip-vit-large-patch14")
+        self.safety_checker = StableDiffusionSafetyChecker.from_pretrained("CompVis/stable-diffusion-safety-checker")
+    def __call__(
+        self,
+        input_image: Union[torch.FloatTensor, PIL.Image.Image, List[PIL.Image.Image]],
+        height: Optional[int] = 512,
+        width: Optional[int] = 512,
+        num_inference_steps: Optional[int] = 50,
+        guidance_scale: Optional[float] = 7.5,
+        eta: Optional[float] = 0.0,
+        latents: Optional[np.ndarray] = None,
+        output_type: Optional[str] = "pil",
+        return_dict: bool = True,
+        **kwargs,
+    ):
+        if "torch_device" in kwargs:
+            device = kwargs.pop("torch_device")
+            warnings.warn(
+                "`torch_device` is deprecated as an input argument to `__call__` and will be removed in v0.3.0."
+                " Consider using `pipe.to(torch_device)` instead."
+            )
+
+            # Set device as before (to be removed in 0.3.0)
+            if device is None:
+                device = "cuda" if torch.cuda.is_available() else "cpu"
+            self.to(device)
+
+        if isinstance(input_image, PIL.Image.Image):
+            batch_size = 1
+        elif isinstance(input_image, list):
+            batch_size = len(input_image)
+        else:
+            raise ValueError(f"`input_image` has to be of type `str` or `list` but is {type(input_image)}")
+
+        if height % 8 != 0 or width % 8 != 0:
+            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")
+
+        if not isinstance(input_image, torch.FloatTensor):
+            input_image = self.feature_extractor(images=input_image, return_tensors="pt").to(self.device)
+
+        image_encoder_output = self.image_encoder.vision_model(input_image["pixel_values"])[1]
+        image_embeddings = self.image_encoder.visual_projection(image_encoder_output)
+        image_embeddings = image_embeddings.unsqueeze(1)
+
+        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
+        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
+        # corresponds to doing no classifier free guidance.
+        do_classifier_free_guidance = guidance_scale > 1.0
+        # get unconditional embeddings for classifier free guidance
+        if do_classifier_free_guidance:
+            uncond_embeddings = torch.zeros_like(image_embeddings)
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            image_embeddings = torch.cat([uncond_embeddings, image_embeddings])
+
+
+        image_embeddings = image_embeddings.cpu().detach().numpy()
+        # get the initial random noise unless the user supplied it
+        latents_shape = (batch_size, 4, height // 8, width // 8)
+        if latents is None:
+            latents = np.random.randn(*latents_shape).astype(np.float32)
+        elif latents.shape != latents_shape:
+            raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {latents_shape}")
+
+
+        # set timesteps
+        accepts_offset = "offset" in set(inspect.signature(self.scheduler.set_timesteps).parameters.keys())
+        extra_set_kwargs = {}
+        if accepts_offset:
+            extra_set_kwargs["offset"] = 1
+
+        self.scheduler.set_timesteps(num_inference_steps, **extra_set_kwargs)
+
+        # if we use LMSDiscreteScheduler, let's make sure latents are mulitplied by sigmas
+        if isinstance(self.scheduler, LMSDiscreteScheduler):
+            latents = latents * self.scheduler.sigmas[0]
+
+        # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
+        # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
+        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
+        # and should be between [0, 1]
+        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        extra_step_kwargs = {}
+        if accepts_eta:
+            extra_step_kwargs["eta"] = eta
+
+        for i, t in enumerate(self.progress_bar(self.scheduler.timesteps)):
+            # expand the latents if we are doing classifier free guidance
+            latent_model_input = np.concatenate([latents] * 2) if do_classifier_free_guidance else latents
+            if isinstance(self.scheduler, LMSDiscreteScheduler):
+                sigma = self.scheduler.sigmas[i]
+                # the model input needs to be scaled to match the continuous ODE formulation in K-LMS
+                latent_model_input = latent_model_input / ((sigma**2 + 1) ** 0.5)
+
+            # predict the noise residual
+            noise_pred = self.unet(
+                sample=latent_model_input, timestep=np.array([t]), encoder_hidden_states=image_embeddings
+            )
+            noise_pred = noise_pred[0]
+
+            # perform guidance
+            if do_classifier_free_guidance:
+                noise_pred_uncond, noise_pred_text = np.split(noise_pred, 2)
+                noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
+
+            # compute the previous noisy sample x_t -> x_t-1
+            if isinstance(self.scheduler, LMSDiscreteScheduler):
+                latents = self.scheduler.step(noise_pred, i, latents, **extra_step_kwargs).prev_sample
+            else:
+                latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs).prev_sample
+
+        # scale and decode the image latents with vae
+        latents = 1 / 0.18215 * latents
+        image = self.vae_decoder(latent_sample=latents)[0]
+
+        image = np.clip(image / 2 + 0.5, 0, 1)
+        image = image.transpose((0, 2, 3, 1))
+
+        # run safety checker
+        safety_cheker_input = self.feature_extractor(self.numpy_to_pil(image), return_tensors="pt").to(self.device)
+        image, has_nsfw_concept = self.safety_checker(images=image, clip_input=safety_cheker_input.pixel_values)
+
+        if output_type == "pil":
+            image = self.numpy_to_pil(image)
+
+        if not return_dict:
+            return (image, has_nsfw_concept)
+
+        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

scripts/convert_sd_image_checkpoint_to_onnx.py

@@ -0,0 +1,161 @@
+# Copyright 2022 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import argparse
+from pathlib import Path
+
+import torch
+from torch.onnx import export
+
+from lambda_diffusers import StableDiffusionImageEmbedOnnxPipeline
+from diffusers.onnx_utils import OnnxRuntimeModel
+from packaging import version
+
+from lambda_diffusers import StableDiffusionImageEmbedPipeline
+
+from PIL import Image
+import requests
+
+is_torch_less_than_1_11 = version.parse(version.parse(torch.__version__).base_version) < version.parse("1.11")
+
+
+def onnx_export(
+    model,
+    model_args: tuple,
+    output_path: Path,
+    ordered_input_names,
+    output_names,
+    dynamic_axes,
+    opset,
+    use_external_data_format=False,
+):
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    # PyTorch deprecated the `enable_onnx_checker` and `use_external_data_format` arguments in v1.11,
+    # so we check the torch version for backwards compatibility
+    if is_torch_less_than_1_11:
+        export(
+            model,
+            model_args,
+            f=output_path.as_posix(),
+            input_names=ordered_input_names,
+            output_names=output_names,
+            dynamic_axes=dynamic_axes,
+            do_constant_folding=True,
+            use_external_data_format=use_external_data_format,
+            enable_onnx_checker=True,
+            opset_version=opset,
+        )
+    else:
+        export(
+            model,
+            model_args,
+            f=output_path.as_posix(),
+            input_names=ordered_input_names,
+            output_names=output_names,
+            dynamic_axes=dynamic_axes,
+            do_constant_folding=True,
+            opset_version=opset,
+        )
+
+
+@torch.no_grad()
+def convert_models(model_path: str, output_path: str, opset: int):
+    pipeline = StableDiffusionImageEmbedPipeline.from_pretrained(model_path)
+    output_path = Path(output_path)
+
+
+    # UNET
+    onnx_export(
+        pipeline.unet,
+        model_args=(torch.randn(2, 4, 64, 64), torch.LongTensor([0, 1]), torch.randn(2, 77, 768), False),
+        output_path=output_path / "unet" / "model.onnx",
+        ordered_input_names=["sample", "timestep", "encoder_hidden_states", "return_dict"],
+        output_names=["out_sample"],  # has to be different from "sample" for correct tracing
+        dynamic_axes={
+            "sample": {0: "batch", 1: "channels", 2: "height", 3: "width"},
+            "timestep": {0: "batch"},
+            "encoder_hidden_states": {0: "batch", 1: "sequence"},
+        },
+        opset=opset,
+        use_external_data_format=True,  # UNet is > 2GB, so the weights need to be split
+    )
+
+    # VAE ENCODER
+    vae_encoder = pipeline.vae
+    # need to get the raw tensor output (sample) from the encoder
+    vae_encoder.forward = lambda sample, return_dict: vae_encoder.encode(sample, return_dict)[0].sample()
+    onnx_export(
+        vae_encoder,
+        model_args=(torch.randn(1, 3, 512, 512), False),
+        output_path=output_path / "vae_encoder" / "model.onnx",
+        ordered_input_names=["sample", "return_dict"],
+        output_names=["latent_sample"],
+        dynamic_axes={
+            "sample": {0: "batch", 1: "channels", 2: "height", 3: "width"},
+        },
+        opset=opset,
+    )
+
+    # VAE DECODER
+    vae_decoder = pipeline.vae
+    # forward only through the decoder part
+    vae_decoder.forward = vae_encoder.decode
+    onnx_export(
+        vae_decoder,
+        model_args=(torch.randn(1, 4, 64, 64), False),
+        output_path=output_path / "vae_decoder" / "model.onnx",
+        ordered_input_names=["latent_sample", "return_dict"],
+        output_names=["sample"],
+        dynamic_axes={
+            "latent_sample": {0: "batch", 1: "channels", 2: "height", 3: "width"},
+        },
+        opset=opset,
+    )
+
+
+    onnx_pipeline = StableDiffusionImageEmbedOnnxPipeline(
+        vae_decoder=OnnxRuntimeModel.from_pretrained(output_path / "vae_decoder"),
+        unet=OnnxRuntimeModel.from_pretrained(output_path / "unet"),
+        scheduler=pipeline.scheduler,
+    )
+
+    onnx_pipeline.save_pretrained(output_path)
+    print("ONNX pipeline saved to", output_path)
+
+    _ = StableDiffusionImageEmbedOnnxPipeline.from_pretrained(output_path, provider="CPUExecutionProvider")
+    print("ONNX pipeline is loadable")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        "--model_path",
+        type=str,
+        required=True,
+        help="Path to the `diffusers` checkpoint to convert (either a local directory or on the Hub).",
+    )
+
+    parser.add_argument("--output_path", type=str, required=True, help="Path to the output model.")
+
+    parser.add_argument(
+        "--opset",
+        default=14,
+        type=str,
+        help="The version of the ONNX operator set to use.",
+    )
+
+    args = parser.parse_args()
+
+    convert_models(args.model_path, args.output_path, args.opset)